[Mesa-dev] [PATCH] nir: Implement optional b2f->iand lowering

[Matt Turner]

On Sun, Apr 29, 2018 at 11:19 AM, Alyssa Rosenzweig
<alyssa at rosenzweig.io> wrote:
> This pass is required by the Midgard compiler; our instruction set uses
> NIR-style booleans (~0 for true) but lacks a dedicated b2f instruction.
> Normally, this lowering pass would be implemented in a backend-specific
> algebraic pass, but this conflicts with the existing iand->b2f pass in
> nir_opt_algebraic.py, hanging the compiler. This patch thus makes the
> existing pass optional (default on -- all other backends should remain
> unaffected), adding an optional pass for lowering the opposite
> direction.

Tell me a little more about what your hardware supports. Does it have
real integers?

In the i965 driver, we used to do what you suggest (AND with
0x3f800000), but changed to a type converting MOV with a source
modifier: MOV dest:F, -src0:D since we could occasionally optimize
that instruction away. That is, interpret the source as an integer (so
false is 0 and true is -1), negate it, and convert to a float.

We did both of these things in our backend when translating NIR's b2f
into our backend IR. Why is it better for you to do that in NIR?